ERCOT Data Tells The Story
There were several headlines recently describing a 25% drop in network difficulty during the mid-January winter storm last week. Most attributed this drop to curtailment activity in Texas. While Texas does represent 17% of the global bitcoin hashrate, ERCOT data shows that some of the curtailment activity was a combination of higher prices and “good grid citizenship.” In ERCOT, and to a lesser extent in other ISOs, prices are the best proxy for grid stress. There are other proxies such as PRC (physical responsive capability) but prices are a better measure for most situations. For that reason, in order to prevent swings in prices and create more challenging gird conditions, an optimal environment is one in which the price does not swing wildly up and down. But that is traditionally what happens (see the graph below from last year’s Winter Storm Elliot). Bitcoin miners are the economically perfect consumers of electricity. That is not to say that bitcoin miners will consume electricity in an altruistic way, but rather that bitcoin miners are so sensitive to the price of power that they are economically incentivized to curtail their consumption when power prices rise past their breakeven (current breakeven for most miners ranges between $100 and $200 per MWh). That means they will be on when prices are below their breakeven and turn off when prices are above it (there are some exceptions to this, for example, if miners are part of a data center that has a hosting agreement with guaranteed uptime).
Texans should want bitcoin miners to be on anytime power is abundant because their presence incentivizes the buildout of additional generation. And less counterintuitively, we naturally want bitcoin miners to curtail when prices are high and the grid is under stress.
That brings us to the January 2024 winter event of the week of January 15th. The headlines would have you think that the Texas grid was again stressed and that bitcoin miners curtailed as a result. The truth is much more nuanced. The average settlement price of power on the wholesale market in ERCOT during the worst three days of the storm was $100.76 per MWh and prices never went over $600 per MWh. For context prices max out at $5,000 per MWh. On the whole, the grid weathered the storm quite well with ample reserves throughout. There was a conservation alert but that was more of a precautionary message for power consumers who don’t monitor the power price every second of every day like bitcoin miners do.
We did see some economic curtailment from miners for extended periods and some for short periods when the prices exceeded $200 MWh, but on the whole, this activity was less pronounced than in previous winter events or summer heat waves. Some bitcoin miners likely curtailed for longer periods in order to be good “grid citizens” and to show their commitment to a stable grid but that is hardly quantifiable. All of this evidence points to a more nuanced understanding of the difficulty drop last week. Much of it was a result of curtailment in Texas but watching the ERCOT price data, leads me to believe that a material portion of that curtailment came from other ISOs in North America as well. In short, everyone who has an opinion about bitcoin mining curtailment would do well to watch ERCOT settlement and LMP prices. The data and the economics should form the backbone of all future analyses.
